We studied the relationship between exposure to malaria, the use of long-term chemoprophylaxis with chloroquine, and the prevalence of sporozoite antibodies in 446 expatriates who had lived in 7 West African countries for 6 months-41 years. Filter paper blood samples from 12% of the subjects had antibodies to the repeat region of the Plasmodium falciparum circumsporozoite protein, with a positive correlation between enzyme-linked immunosorbent assay (ELISA) absorbance and years of exposure (r = 0.32, P = < 0.01). Development of sporozoite antibodies did not correlate with reported use of chloroquine. Ten samples from expatriates with the highest ELISA titers and 10 samples from West African nationals, which were matched for ELISA titer and duration of exposure, were characterized more fully. All 20 samples reacted strongly with sporozoites by immunofluorescence. The 10 samples from nationals reacted strongly with liver-stage antigens by immunofluorescence and with blood-stage antigens by immunofluorescence and immunoblotting. In contrast, the 10 samples from expatriates were negative or only weakly positive in the liver- and blood-stage assays. These results imply that sporozoite antibodies are generally not cross-reactive with blood-stage antigens, and suggest that protective immunity to malaria does not develop during long-term malaria chemoprophylaxis against the erythrocytic stage.